PWM article
jsn
conversion this morning. Here is a really cool article that describes
how these things work - although from the standpoint of creating PWM
output from a microcontroller:
http://www.evilmadscientist.com/article.php/avrdac
Plus it's from Evil Mad Scientist labs so you know it's gotta be good :)
jsn
jsn
http://sound.westhost.com/articles/pwm.htm
So would it be possible to build an amplifier that accepts input as
digital data (from some sort of digital output on a CD
player/soundcard/squeezebox) and never really convert it to analog
until the PWM output to the speaker? That seems like an elegant (if
inflexible) design that would eliminate a lot of the D to A, back to
D, back to A steps when you use a PWM amp with a digital source.
jsn
John Galbraith
distortion are just bound to be worse than a traditional amp, either class A
or AB. I'm sure the waveform "looks" right, but given the effort we all put
into traditional amps and their power supplies, it just sounds crazy to accept
that level of distortion right in the audio path.
My second huge complaint is that it requires a very fast circuit to sound
good. If I've got parts with a 1MHz bandwidth or gain-bandwidth product, I
want to exploit that for a nice high end with good phase response, not
synthesize a lot of noise of there that has to be filtered. Remember that
phase effects are heard a whole decade under the cutoff frequency, so it's nice
to have at least 200 kHz of bandwidth even in a traditional amp. It's hard to
make a fast circuit, and one can only do so well. It seems silly to waste the
bandwidth this way.
Now there are two huge advantages that most DIY'ers just don't care about:
cost and weight. Granted, the class D amp will win here big time.
I would use class D only on very low frequency, big power applications like a
subwoofer, or else low-fi applications like sound reinforcement where you want
tons of power but also have to move heavy gear when nobody really cares how
good it sounds anyway.
John
jsn
this context?
And yeah, I agree that it seems nuts that it would even work, much
less sound good.
I like them for the cheapness, and because you can then get several
channels of audio to do bi-amping. I haven't done much with this, but
I like the concept :)
Also, the cost and weight would be nice for PA applications.
jsn
> --
> You received this message because you are subscribed to the Google Groups "audionerdz" group.
> To post to this group, send email to audio...@googlegroups.com.
> To unsubscribe from this group, send email to audionerdz+...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/audionerdz?hl=en.
>
>
John Galbraith
> 200KHz bandwidth seems really excessive. What are phase effects in
> this context?
Well, it's obsessive for sure, but the phase effects are delay. So if you
really wanted absolutely perfect transient response and great imaging at 20kHz
you would want the first pole of the amplifier to be 200kHz or over.
You can probably make a pretty decent sounding amp with less bandwidth. 8-)
John
jsn
If you measure a really amazing tweeter, can you see response up to 100kHz?
jsn
John Galbraith
> Oh I get it - like in an "impulse response" sort of way. Pretty cool :)
>
> If you measure a really amazing tweeter, can you see response up to 100kHz?
Absolutely. It may be tens of dB down, but you could totally measure it.
Most network analyzers have easily 70-80 dB dynamic range or more. Also, I
think that SACD might actually reproduce frequencies up there, and DVDA is
close (it's 192 kHz sample rate?) So not only can one measure it, but we
should all strive to achieve it in our projects. 8-) The network analyzer
will also measure the delay.
John